رشد، توزیع درآمد و نوسانات سیاست مالی

The relationship between income distribution and economic growth has long been an important economic research subject. Despite substantial evidence on the negative impact on long-term growth of inequality in the literature, however, there is not much consensus on the specific channels through which inequality affects growth. The empirical validity of two most prominent political economy channels – redistributive fiscal spending and taxes, and sociopolitical instability – has recently been challenged. We advance a new political economy channel for the negative link between inequality and growth, a fiscal policy volatility channel, and present strong supporting econometric evidence in a large sample of countries over the period of 1960–2000. Our finding also sheds light on another commonly observed negative relation between macroeconomic volatility and growth. We carefully address the robustness of the results in terms of data, estimation methods, outlier problem, and endogeneity problem that often plague the standard OLS (ordinary least squares) regression.

The relationship between income distribution and growth has long been an important economic research subject. The earlier development literature argued that high inequality could help growth by directing scarce resources to high-saving capitalists (Lewis, 1955 and Kaldor, 1956). With the advent of endogenous growth theories which made it possible to derive a theoretical relation between inequality and growth, economists renewed their interest in the age-old question. The new growth literature has reversed the earlier prediction and provided persuasive intellectual support for the proposition that inequality harms growth through political economy channels (see Alesina and Rodrik, 1994, Persson and Tabellini, 1994 and Benabou, 1996 for seminal work, and Drazen, 2000 for a literature survey). Overall, there is substantial supporting evidence on the negative relation between inequality and long-term growth (Easterly, 2007, Perotti, 1996, Clarke, 1995 and Birdsall et al., 1995 for example). Yet there is not much consensus on the specific channel(s) through which inequality affects growth negatively. Some of the earlier empirical studies have focused on the reduced-form relationship between inequality and growth, and hence cannot shed light on the underlying mechanisms that may have different policy implications. Moreover, the empirical validity of two most prominent political economy channels – redistributive fiscal policy (Alesina and Rodrik, 1994 and Persson and Tabellini, 1994), and political violence and instability (Benhabib and Rustichini, 1996 and Alesina and Perotti, 1996) – has recently been challenged (see Perotti, 1996, Keefer and Knack, 2002a, Campos and Nugent, 2002, Mulligan et al., 2004 and Glaeser, 2006). That is, there is little evidence that high inequality leads to more redistribution or political instability.1 This paper empirically investigates the relationship between income distribution, fiscal policy, and growth with multiple purposes. We make a contribution to the literature by offering an empirical analysis of a new mechanism for the negative link between inequality and growth: a fiscal policy volatility channel. Intuitively, struggles over income distribution in highly unequal societies can lead to sharp disagreements over ideal government policies. In the presence of social preference polarization, heterogeneous policymakers may have greater incentives to insist on their preferred policies. Such incentives to put forward their preferred agenda may become particularly strong during good times when rising government revenues or newly available resources make their agenda seem more feasible, which results in a pro-cyclical fiscal spending increase and hence a fiscal deficit (especially when institutional constraints are insufficient). At the same time, discretionary spending decisions taken in such a manner yield volatile fiscal outcomes over time. In a non-cooperative fiscal game model embedded in an endogenous growth framework, Woo (2005) formalizes this intuition and shows that in a highly polarized country (due to high inequality), discretionary fiscal spending becomes procyclical and volatile, which in turn reduces growth along the transition path towards a new lower output.2 This is a new fiscal policy channel that is distinct from the aforementioned redistribution channel. Yet there is no empirical analysis on this channel of fiscal policy volatility. In this paper, we fill this gap in the literature. It is important to distinguish fiscal volatility from adaptability (or flexibility) to sudden changes of economic conditions, such as counter-cyclical fiscal response to macroeconomic shocks, because the latter is more likely to stabilize the economy and promote economic growth, rather than discourage growth. In our theory, the notion of fiscal policy volatility corresponds to excessive discretionary changes in fiscal policy that take place for reasons other than smoothing out output fluctuations or responding to macroeconomic conditions.3 Fiscal procyclicality has a parallel with fiscal volatility. Unlike the fiscal response to macroeconomic conditions, discretionary and procyclical fiscal changes can be harmful to growth by amplifying economic volatility.4 Accordingly, we focus on the empirical counterpart of this notion of fiscal policy volatility, and present strong evidence that countries with high initial income inequality tend to suffer greater fiscal policy volatility and procyclicality, which in turn hinders economic growth in a large sample of countries over the period of 1960–2000. Also, we explore the channels through which fiscal volatility and procyclicality influences growth in income per worker by combining the growth accounting decomposition with regression analysis. We find that the negative effects on output per worker growth from fiscal policy volatility and procyclicality are mainly through the channels of TFP and capital per worker growth. Our main results are robust to various estimation techniques, alternative measures of fiscal policy volatility and cyclicality, and different sub-sample periods (1960–2000, 1970–2000, and 1980–2000). Indeed, we carefully address the robustness of our results in terms of data, estimation methods, outlier problem, and endogeneity problem that often plague the standard OLS (ordinary least squares) regression analysis. Our paper is related to existing studies in various aspects. First of all, income inequality has long been mentioned as an important factor in understanding why populist fiscal policies appear more often in Latin American countries than other regions such as East Asia, and also in explaining contrasting macroeconomic performance across developing regions (Rodrik, 1996, Birdsall et al., 1995 and Sachs, 1989 among others). In a well-known study, Engerman and Sokoloff, 2002 and Engerman and Sokoloff, 2005 make historical observations that initial differences in the extent of inequality across countries contributed to the different decisions they made regarding the types and size of government expenditure programs, how much revenue to raise, and the relative use of different tax instruments, which ultimately influenced long-run paths of development. However, the existing empirical studies have focused on the link between income distribution and other macroeconomic outcomes such as external debt (Berg and Sachs, 1988), public deficits (Woo, 2003) and inflation (Desai et al., 2005). Yet they do not attempt to establish a channel between inequality and growth. Moreover, the existing evidence on the negative effects on growth of these macroeconomic imbalances is at best mixed. In a recent evaluation of growth regressions in relation to macroeconomic variables, Easterly (2005) finds that some of the large effects of a policy variable(s) such as inflation and budget deficits on growth are often caused by outliers that represent “extremely bad policies”. In this paper, we focus on fiscal policy volatility in relation to inequality and examine its link to economic growth. There is a large and growing political economy literature that has studied the effects on fiscal policy of political variables such as political regime, political instability, electoral rules, party structure, institutionalized constraints including veto players (see Milesi-Ferretti et al., 2002, Persson and Tabellini, 2003, Persson and Tabellini, 2006 and Scartascini, 2007 for example). Most of the existing studies have studied government budget deficit, expenditure and spending composition, rather than fiscal volatility/procyclicality per se. Thus, some of the theoretical implications from this political economy literature are not always directly applicable to our issue at hand. Nonetheless, there is substantial evidence that fiscal outcomes are heavily influenced by political factors, and the issue of fiscal volatility and procyclicality may not be an exception. Intuitively, the effects on fiscal volatility and procyclicality of inequality may be even more pronounced or suppressed, depending on political and institutional structures that influence the fiscal policy-making process. For example, institutional constraints that function as a checks and balances mechanism may discourage large discretionary fiscal policy changes, contributing to lower fiscal volatility. Also, the checks and balances may act as a consensus-building institution and work to suppress conflicts of interest among different groups, making the social polarization effect from inequality less important in determining the fiscal outcomes. Closely related to this are veto players. Multiple veto players may contribute to reducing the fiscal discretion of the incumbent government and avoid abrupt large changes in fiscal policy by providing checks and balances in the fiscal decision-making process. At the same, the existence of multiple veto players may simply mean a status quo, contributing to delaying the necessary fiscal adjustments which would require eventual bigger fiscal restructuring. In this case, multiple veto players may actually be associated with larger fiscal volatility over a time period, which makes the effects of veto players somewhat ambiguous a priori. However, constraints on executive decision makers, multiple veto players, and the political institutions that tend to produce multiple veto players turn out to be associated with the smaller magnitude of fiscal policy volatility.5 Interestingly, initial inequality continues to be statistically significant in explaining cross-country variation in fiscal volatility and procyclicality, even after controlling for measures of institutional constraints or political variables. Our finding on the link between inequality and fiscal volatility may also help explain another important empirical regularity: a negative relation between macroeconomic volatility (as measured by volatility of real GDP) and growth (Ramey and Ramey, 1995 and Fatas and Mihov, 2003). Our paper suggests that macroeconomic volatility may reflect government policy-induced volatility which itself is rooted in initial income inequality. Therefore, the negative linkage between macroeconomic volatility and growth may reflect a part of the chain between initial inequality and growth via the fiscal policy volatility channel which our paper tries to highlight. The plan of the paper is as follows. Section 2 presents reduced-form growth regressions on income inequality. 3 and 4 separately examine the two links in the fiscal policy volatility mechanism: link between initial inequality and fiscal volatility, on the one hand, and link between fiscal volatility and growth, on the other hand. A range of political and institutional variables are also considered along with the relevant political economy arguments. Various robustness checks follow. Concluding remarks are in Section 5. We provide detailed information on data including the country list, sources and summary statistics in the data appendix.

We examined a new channel of fiscal policy volatility/procyclicality that relates initial income inequality to poor growth in a cross-country data over the period of 1960–2000. Not only could we confirm the negative relation between initial income inequality and long-term growth in our sample, but also presented the first econometric evidence for both the relation between initial inequality and fiscal volatility/procyclicality, and the relation between fiscal volatility and growth. Moreover, we carefully checked robustness of our results by addressing the problems of outliers and endogeneity. In short, the evidence consistently supports our hypothesis that more unequal societies are more likely to resort to more aggressive use of discretionary and procyclical fiscal policies that are detrimental to economic growth. Also, we considered a range of political variables in four particular aspects—political instability, government fragmentation, political regime and electoral system, and institutional constraints. The econometric results on the political factors in explaining cross-country variation of fiscal volatility and procyclicality are mixed. Nonetheless, there is evidence that suggests a potential role of institutional constraints on policymakers in reducing the discretionary fiscal policy volatility and procyclicality. Yet it remains an important question how we can make the institutional constraints work to maintain stable fiscal policies in a country with high inequality. In other words, one can ask under what conditions and what types of institutional constraints on fiscal policymakers work (or do not work) in a country with highly unequal societies where the social preference over fiscal policies may be sharply divided. This is an interesting and important topic for research in the future.